Rigol HDO1000 and HDO4000 12bit oscilloscopes launched in China

[vokars]

The command "%matplotlib inline"  is not working on mine(SyntaxError: invalid syntax)

This is a so-called Jupyter magic and not a Python command, but is interpreted by the Jupyter notebook: https://ipython.readthedocs.io/en/stable/interactive/magics.html
With the alternative %matplotlib notebook you can scale and zoom the diagram.

The two Bode plots show an approximately constant voltage. If you comment out the model part (delete the """ lines), the function generator produces a frequency dependent voltage and in this way emulates the DUT.

Manual scaling of the y-axis can be done with plt.ylim(<low>,<high>). With a Jupyter notebook you can easily copy the Matplotlib / diagram part to another cell and perform additional analysis as all measurement data is still stored in the result lists (f_log_list, ...).

Typical result with function generator emulation attached.

[skander36]

The command "%matplotlib inline"  is not working on mine(SyntaxError: invalid syntax)

This is a so-called Jupyter magic and not a Python command, but is interpreted by the Jupyter notebook: https://ipython.readthedocs.io/en/stable/interactive/magics.html
With the alternative %matplotlib notebook you can scale and zoom the diagram.

The two Bode plots show an approximately constant voltage. If you comment out the model part (delete the """ lines), the function generator produces a frequency dependent voltage and in this way emulates the DUT.

Manual scaling of the y-axis can be done with plt.ylim(<low>,<high>). With a Jupyter notebook you can easily copy the Matplotlib / diagram part to another cell and perform additional analysis as all measurement data is still stored in the result lists (f_log_list, ...).

Ok. Thank you for your answer. My reply was just a feedback. I saw it DUT emulation part, but I preferred to use a simple RC filter.
 

[Frunse]

Hi the Script is running but it is working not right i guess or i make something wrong.

My Generator is switching the Frequencies and gives 2VPP Out to the Scope on Channel One with 50Ohm Termination
Stopp Frequency is 10MHz

The Scope gives me after each reading on all 4Channels (all Channels goes on for short Time) this Error Massage:

Remote cmd execute error!

And the Bode Plot looks Strange / wrong.

Robert

[Frunse]

Hi again,
found the Error

Was the thing in between my Ears called Head

On Generator i was @ HighZ and not @ 50Ohm
Now it works also without the Error Massage on Scope.

On Script i changed only this:

fg=rm.open_resource('TCPIP0::192.168.0.103::INSTR')

scope=rm.open_resource('TCPIP::192.168.0.95::INSTR')

# function generator set frequency
lc_str = "C1:BSWV FRQ," + str(lc_f_Hz)

measured an simple RC Filter -3dB Point @ around 3kHz
with 1k and 0,1uF

Robert 

[vokars]

measured an simple RC Filter -3dB Point @ around 3kHz
with 1k and 0,1uF

 

Now you could increase the resolution by increasing gl_points_per_decade  .

[Frunse]

Yeah did it

Very nice tool, and again thank you for youre Work and for Sharing it with us noobs

Robert

[vokars]

 

[Frunse]

So if i am thinking right,

also possible to make an  Curvetracer this way to masure Transistors or diodes ?

With an nice Plot of the measuring Curves ?

The XY Mode of an Scope does this job too but i would realy like to have an nice Bodeplot of the Curves with right scale like on Transistor datasheets.

Possible with Python?

Robert

[vokars]

The XY Mode of an Scope does this job too but i would realy like to have an nice Bodeplot of the Curves with right scale like on Transistor datasheets.

Possible with Python?

It should not be a problem. Python is very flexible. You could use the Bode Plot blueprint and expand a second channel (would be necessary for precise measurement anyway). Measure current with shunt resistor. Instead of frequency here voltage has to be controlled in the signal source. Apply DC voltage with a remote controllable power supply or, as a workaround, use a function generator with square wave signal and low frequency (< 0.1 Hz, lower than the measurement time of the oscilloscope, choose phase appropriately) if the function generator cannot generate a simple DC voltage.

And of course it is possible to apply a sawtooth from the function generator and measure the current with a shunt resistor. The trigger is external. In this case you will need another script that can download the two waveforms from the oscilloscope. The sampling rate should not be too high for this so that you don't need too much memory. The sampling rate in Rigol can be reduced indirectly by reducing the memory. The formula for this is in the user manual.

[Frunse]

Thanks,
i will go into it after i done my running Projects, seems need some time to deal with it.

But will be realy fun to play with the Instruments and Python, also have an nice UDP3305S to use for this

Robert

[vokars]

also have an nice UDP3305S to use for this

Then it should be straight forward to use the scan method like the Bode Plot script does.

[Frunse]

Found this Video very good!



Easy-SCPI  could made all easier i guess

Need to go into it.

Robert

[Frunse]

Hi,
have tested the DSO1074 about the noise Values:

Siglent has only Time base Accuracy ±25ppm on Datasheet Page 9 on bottom you can read it.
https://www.batronix.com/files/Siglent/Oszilloskope/SDS1000X-HD/SDS1000X-HD_Datasheet_EN01E.pdf

Siglent writes on his Datasheet:
Noise floor*2 (rms, @50 Ω, typical,1 mV/div) 70 μV(Full Bandwith)

With 50Ohm Termination i measured 63uV @full BW and
23uV @ 20MHz BW Limit AC RMS on Rigol DHO1074(hacked)

Robert

[ebastler]

Siglent has only Time base Accuracy ±25ppm on Datasheet Page 9 on bottom you can read it.
https://www.batronix.com/files/Siglent/Oszilloskope/SDS1000X-HD/SDS1000X-HD_Datasheet_EN01E.pdf

Siglent writes on his Datasheet:
Noise floor*2 (rms, @50 Ω, typical,1 mV/div) 70 μV(Full Bandwith)

With 50Ohm Termination i measured 63uV @full BW and
23uV @ 20MHz BW Limit AC RMS on Rigol DHO1074(hacked)

The time base accuracy is indeed a strength of the Rigol DHO1000 -- and you can even connect an external 10 MHz reference if the time base accuracy should really be critical for your application.

The noise floor is not exceptional though. Actual values measured on my SDS814X HD (also hacked to 200 MHz nominal bandwidth, higher in practice) are somewhat better than what you found for the Rigol. I used the same values you chose, although I would normally go to a slower time base to capture the 1/f noise better.

[Frunse]

Hi,
youre scope is by sure an good Scope, Price Ratio is also great, but i never will buy and Scope with an smal Display and without HDMI Out anymore.

I would struggle with my eyes using it, like on my older DS1054Z.

Can't see on youre picture if you are using AC_RMS Measuring is this normal on Siglent???

Rigol says low noise to 18uVrms
but i am not sure how i could reach it, also don't find the settings they use for that may be anyone could give this Information?

But i am fine with it, how it is, It has all i want from an good 12Bit Scope

And if Rigol will aktivate the 50 Ohm Option any day because all Parts soldered for this already (evvblog tear down) that would be even greater (a little wisch)
Not sure if they ever has a plan to do it thou.

Found this Information on the Siglent Datasheet of youre Scope, look at the attached Picture.
And made the same Measuring with my Scope with same Values for Time Base and V/Div.
But i need to use Memory 10MPoints for 2Gs/s no 2MPoints possible and if i use 1MP he switched to 1Gs/s
I have Hardware Version 2 and read there is already Hardware version 3 out, would be also interesting if anything changed on Noise Performance.

Next Question is:
Would an Mains Filter help in front of an Scope for better Noise Performance or that all comes from the Quality of that used ADC. ?
Need to check this too next

So Youre Measured Values are a lot Better than on the Datasheet, you lucky Dude

Robert

[ebastler]

youre scope is by sure an good Scope, Price Ratio is also great, but i never will buy and Scope with an smal Display and without HDMI Out anymore.
[...]
Can't see on youre picture if you are using AC_RMS Measuring is this normal on Siglent???
[...]
So Youre Measured Values are a lot Better than on the Datasheet, you lucky Dude

All good, it's not meant to be a competition. I just wanted to point out that, while the time base accuracy of the DHO1000 is exceptional for its class, the noise is par for the course. Which is still pretty good, and certainly a major improvement over any typical 8-bit scope. 

The display is clearly another DHO1000 strength. I had a DHO1000 in December/January, and while am still happy with my decision to switch to the Siglent, I would not mind getting the larger, high-res screen back!

Siglent uses the term "StdDev" for what Rigol calls "AC.RMS". One company using mathematical terminology, one the EE term, but both mean the same thing. My screenshots show standard deviation measured in all four channels. In addition I have enabled statistics, so you best look at the "Mean" line, which shows the StdDev averaged over many sweeps for a more stable readout.

It indeed seems that either Siglent was very conservative in specifying the noise floor, or I have a better-than-average unit. I just checked the single-channel noise at 500 µV/div and 100 µs/div and get just below 50 µV (rms or std. deviation value). And my scope still thinks it's an SDS824X HD. Not sure what's going on, but I won't complain... 

[Frunse]

Hi,

sure not an Compilation, but i always like to see how other Scopes perform in Comparsion to his Datasheets.

Also would realy like to see this kind of stuff from the SDS1000X HD Series in Comparsion.

Also better on Siglent is the FFT i guess he using 2Mpoints instead of 1MPoints on Rigol, right?

For That i am using the QA403 for Audio and the TinySA Ultra for Higher RF Values.
May be any day, if i got lots of Money (Dreaming LOL) an Spektrumanalysator in Kombination with an VNA would be nice Toy but they are so damm expensive
To Expensive for Hobby Purpose.

Robert

[ebastler]

Also would realy like to see this kind of stuff from the SDS1000X HD Series in Comparsion.

The SDS1000X HD is quite similar to the 800X HD, the larger screen (but with the same resolution) being the most obvious difference. Performance data are exactly the same as for the 824X HD, if I have not overlooked anything, and the internal architecture is the same. But there are a few additional features, notably including 50 Ohm inputs.

Also better on Siglent is the FFT i guess he using 2Mpoints instead of 1MPoints on Rigol, right?

It's not so much the larger number of data points, but the more thoughtful software implementation, I'd say. E.g. you get Averaging and Peak Hold modes for the FFT, which Rigol does not offer. In general, the maturity of the software was what tipped the scale in favor of the SDS800X HD for me.

But this is drifting off-topic. If you want to learn more about the Siglent's capabilities, there is a great in-depth review & demonstration thread by forum member Performa01.

[Kurt_!]

A picture of my DHO
50Ohm

[ebastler]

A picture of my DHO
50Ohm

Seems to be broken. 

[Kurt_!]

One more thing

[Kurt_!]

The right settings

[ebastler]

The right settings

Yep, averaging helps. (Seriously, it does -- works around the noise to make best use of the 12 bit resolution. For non-periodic signals, applying a low-pass filter via a Math function is also worth considering.)

Stopping the acquisition and then changing settings before the screenshot is considered bad form though...

[Njk]

The HDO1000U serie has released recently. It's slightly cheaper, uses the same optical encoders and the spec says it has twice less memory. Is it the only limitation?

[ebastler]

The HDO1000U serie has released recently. It's slightly cheaper, uses the same optical encoders and the spec says it has twice less memory. Is it the only limitation?

Unless the -U series specs have changed, it also has a slower waveform capture rate. 30 k/s normally (vs. 50k/s for the regular DHO1000), and 500 k/s in the not-so-useful UltraAcquire mode (vs. 1,500 k/s).